Rigid-Flex Circuits Resources
If you are not used to designing with flex circuits, designing your first rigid-flex can be hard. Just like with flex, it’s as much about the mechanical design of the circuit as it is about the electrical implications of it. But with some research, careful designing, and following the best rigid-flex practices, you will be able to design an amazing rigid-flex, maybe even without needing the paper doll method.
Rigid-Flex Circuit Resources
When designing a rigid-flex, you have to consider a lot of things. The very first is cost. Rigid-flex is hard to fabricate, and if the cost is your primary concern, maybe the same design can be realized in a Ridized-flex circuit (flex circuit using stiffeners). Also, where will you be using the circuit? For high-speed and high-frequency applications, you may have to go with specialized materials like Megtron 6, instead of the typical FR-4. Another major issue will be whether to use any components on the flex part of the circuit or use it primarily to connect two rigid sections of your PCB.
All these and many more considerations go into designing a successful rigid-flex.
Normally, a rigid-flex is designed using FR-4 for the rigid parts of the circuit and PI for the flex part. PET is a cheaper alternative to PI, but only if your design can be fabricated in low-heat conditions since the PET isn’t very heat-resilient. For the rigid part, FR-4 is the most cost-effective selection. But for certain applications, especially high-frequency HDI, you may need to design with a different material in mind.
But it’s important to understand that stacking up a bunch of different substrate materials, even just on the rigid side, requires a lot of care. You have to make sure that the thermal properties (like co-efficient of thermal expansion CTE and Tg) of the substrates align with each other and with the conductor (copper). Because if they don’t, the copper might expand beyond the feasible limits, or the circuit might delaminate.
The copper for flex can be the typical electrodeposited laminated copper foil if you are designing a static flex (bent once). But for dynamic application where the flex has to go through many bending cycles, use rolled annealed foils.
Adhesiveness PI is usually the best choice. It keeps the flex thickness to a minimum.
Some design considerations will help you with the design of a cost-effective and functional rigid-flex.
- If you can avoid it, don’t place any components on the flex part of the circuit. This will allow for a higher degree of flexibility and will keep the flex length to a minimum (since you wouldn’t need to reconcile component placement and bend radius of the flex).
- For a dynamic rigid-flex, keep the flex layer count to a minimum and, if possible, opt for hexed polygons in the signal and ground layers for higher flexibility.
- Using a table to define a stack layer design will help you identify the rigid and flex areas of the circuit. This will help you clean out the transition points.
- Don’t place components near the transition point (where rigid meets flex). Keep them at least 20 mils away.
- Don’t abruptly change the widths of copper traces on the flex. It creates a weak point and increases the possibility of a “fracture” in the copper. Gradually thin the traces instead. Also, if you are going for double-sided flex routing for a relatively complex design, don’t create an I-beam effect. Alternate the traces between the top and bottom layers (or adjacent inner layers if you have a multilayer flex).
- If you have to place components on the flex part, avoid the bend lines altogether. For better stress management, plate the through holes and anchor your surface-mounted pads by additional coverlay. Also, keep the plated through holes as far away from the traces as possible.
- Avoid sharp turns in the flex. Calculate the bending radii as per the number of flex layers. When in a multilayer flex, different layers are mechanically separated with an air gap between them, bending at a larger angle can cause compressions in the inner layer. To avoid this issue, you should design using the “book-binding” method (different length flex layers)
- Use tear-drop formation on the transition points. A sharp disconnection from the rigid-to-flex area of the circuit can cause undue stress on the copper layer near that area.
- Always try to go with a homogenous rigid-flex design. Even the number of rigid layers, the same number of layers in all the rigid areas of the circuit, and flex sandwiched between equal numbers of rigid layers.
- The traces should always be perpendicular to the bending line. And when the flex is in a curved shape, make sure the traces are curved as well. Sharp turns in the traces to prevent fatiguing in the copper.
Design Software and Format
When designing a rigid-flex, make sure the design files are in a format supported by your fabricator. High-end design software like Altium allows you to realize almost any rigid-flex design. It’s important to separate the flex layers from the rigid layers in the design since they will both be using different stack-ups (the rigid stack-up will usually have a higher number of layers.)
Designing in good software will allow you to understand the Z-axis of your design better and work out efficient component placement in the inner layers. Also, you can see the 3D rendering of your Rigid-Flex, which will give you an idea about its physical dimensions and how it will fit its housing.
Rigid-flex is a powerful technology that more than makes up for its cost factors, with its mechanical and electrical superiority over rigid and flex PCBs. And there are a lot of ways to control costs. The best way is to keep all the design considerations in mind, use as simple stack-up as your design allows, and consult your fabricator. Always follow the guidelines of your fabricator.
Working with experienced fabricators will allow you to get the best results for the most reasonable cost.